English In Domains Research Articles

Improving Semantic Dependency Parsing with Higher-Order Information Encoded by Graph Neural Networks

Higher-order information brings significant accuracy gains in semantic dependency parsing. However, modeling higher-order information is non-trivial. Graph neural networks (GNNs) have been demonstrated to be an effective tool for encoding higher-order information in many graph learning tasks. Inspired by the success of GNNs, we investigate improving semantic dependency parsing with higher-order information encoded by multi-layer GNNs. Experiments are conducted on the SemEval 2015 Task 18 dataset in three languages (Chinese, English, and Czech). Compared to the previous state-of-the-art parser, our parser yields 0.3% and 2.2% improvement in average labeled F1-score on English in-domain (ID) and out-of-domain (OOD) test sets, 2.6% improvement on Chinese ID test set, and 2.0% and 1.8% improvement on Czech ID and OOD test sets. Experimental results show that our parser outperforms the previous best one on the SemEval 2015 Task 18 dataset in three languages. The outstanding performance of our parser demonstrates that the higher-order information encoded by GNNs is exceedingly beneficial for improving SDP.

Revisiting the prevalence of English: language use outside the home in South Africa

ABSTRACT In this study, we revisit the status of English relative to the African languages in South Africa by analysing new national data on the main language spoken outside the home. These data, which derive from the General Household Surveys of 2017 and 2018, complement commonly collected data on the main language spoken within the home. Our analysis shows that only a small minority of ‘Africans’ report speaking English most often outside the home, and that the large majority speak the same African language both inside and outside the home. These findings suggest that the dominance of English must be distinguished from its prevalence, and they point to the continued vitality of African languages in the country. In the latter part of the study, we discuss various reasons for these language patterns, including the continuing salience of residential segregation by race, changes in the labour market and the accompanying rise of the African middle-class, and the significance of African languages as markers of identity and resistance to the importance of English in domains of power.

Neural ParsCit: a deep learning-based reference string parser

We present a deep learning approach for the core digital libraries task of parsing bibliographic reference strings. We deploy the state-of-the-art long short-term memory (LSTM) neural network architecture, a variant of a recurrent neural network to capture long-range dependencies in reference strings. We explore word embeddings and character-based word embeddings as an alternative to handcrafted features. We incrementally experiment with features, architectural configurations, and the diversity of the dataset. Our final model is an LSTM-based architecture, which layers a linear chain conditional random field (CRF) over the LSTM output. In extensive experiments in both English in-domain (computer science) and out-of-domain (humanities) test cases, as well as multilingual data, our results show a significant gain ( $$p<0.01$$ ) over the reported state-of-the-art CRF-only-based parser.

The World System of Englishes

Contact between and mutual influences among varieties of standard and non-standard English have always been a central concern in research on World Englishes. In a mobile and globalising world such contacts are by no means restricted to diffusion of features in face-to-face interaction, across contiguous territories in space or up and down the sociolinguistic scale. In order to better represent and understand the complex relationships obtaining between varieties of standard and non-standard English in the contemporary “English language complex” (McArthur 2003: 56; Mesthrie and Bhatt 2008: 1–3), the present paper proposes a new theoretical model, based on language systems theory (de Swaan 2002, 2010). While the model is not designed to supersede existing alternatives, such as the Kachruvian (1982) Circles, it will nevertheless complement them in important ways, chiefly because it is better equipped to handle uses of English in domains beyond the post-colonial nation state. The “World System of Englishes” model was developed in the course of the author’s work on the use of pidgins and creoles in web forums serving the post-colonial West African and Caribbean diasporas. The way Nigerian Pidgin figures in the creation of a globalised digital ethnolinguistic repertoire will hence serve as an illustration of its usefulness.

Truncated Multilingualism or Language Shift? An Examination of Language Use in Intimate Domains in a New Non-racial Working Class Township in South Africa

The dominance of English in particular domains of language use in South Africa, such as higher education and the economy, has led to the fear that other languages may be threatened by an increasing shift to English in all domains, especially among the young. However, this paper reveals the strong vitality of the mother tongues in the intimate domains and increasing multilingualism becoming the norm in new shared spaces.